In developed countries the care of complex older patients commonly entails their transfer from one health care setting to another and from one team of providers to another, as their medical and functional needs evolve. In theory, this allows an economical application of specialized resources to be applied at each stage. As an example (Figure 16-1), an elderly woman may be admitted to an acute hospital for diagnosis and initial management of community-acquired pneumonia. After stabilization and initiation of treatment, she is transferred to a skilled nursing facility (SNF) for rehabilitation, continued antibiotic treatment, and close monitoring by nursing. If she decompensates in some manner in the nursing facility (for example, by becoming acutely delirious), she may be transferred to an emergency department (ED) for urgent physician evaluation; rehospitalization may follow. After further diagnostic work and treatment, she returns to the nursing facility to complete her rehabilitative program before finally returning to her home in the community, where a visiting nursing agency assists family in her care. In this sequence her changing care needs prompted physical transfers and encounters with a succession of professionals—hospitalists and acute care nursing in the hospital; then a new provider, nursing, and rehabilitation staff at the SNF; then new providers and nursing in the ED; probably new hospitalists and nurses on her rehospitalization; and finally a new set of nurses, rehabilitation team, and outpatient provider when she returns home. To keep this patient from falling through the cracks, her various providers (and their organizations) must exercise considerable diligence and attention to detail. Active involvement from the patient and her family is necessary as well.

This chapter discusses transitional care, the actions taken to ensure coordination and continuity of health care as patients are transferred among various care settings. Good transitional care entails not just the physical transfer of a patient, but the orderly transfer of responsibility of care of the patient as well. The chapter is meant to be of particular use for those clinicians (physicians, physician assistants, and nurse practitioners) charged with direct responsibility for managing care transitions.

For a number of reasons, the challenges of transitional care appear to be growing in the United States. Compared to the situation faced by physicians a century ago, our patients are much older, with longer problem (and medication) lists, and greater functional disability. Families may live in other cities and offer little ability or willingness to provide care at home. Health care venues have proliferated and become specialized. Typical care settings found in many large American cities are listed in Table 16-1. Providers themselves have also become more specialized, and in some instances (as with hospitalists, intensivists, SNFists, or emergency medicine physicians), their specialty is defined by the setting in which they practice. Changes in medical education are introducing new challenges as well, with resident work hour limits forcing more handoffs than existed before. Finally, coordination and continuity of care is made more difficult by the fact that the United States lacks a health care “system” per se.

Transitional care is at the confluence of two ascendant movements in modern health care—improved patient safety and patient-centered care. Poorly executed transitions place patients at risk of injury and other adverse events. Well-executed transitions intrinsically incorporate patients’ individual goals, needs, and values. The quality of transitions also influences resource use, such as ED utilization and need for rehospitalization. For all these reasons (Table 16-2), a number of influential advisory, professional, and funding organizations have shown interest in the topic. The Institute of Medicine has called for greater integration of care delivery across health care settings. The Society of Hospital Medicine, the American Geriatrics Society, and the American Board of Internal Medicine all have urged greater professional commitment to transitional care. The Joint Commission on the Accreditation of Healthcare Organizations (JCAHCO) has adopted a patient tracer methodology, which allows surveyors to determine how a patient fares during a sequence of transitions. JCAHCO also has begun confirming that surveyed organizations employ medication reconciliation (discussed below) and standardized approaches for patient handoffs.

Hospitals are required, as part of Medicare Conditions of Participation, to employ a discharge planning process for all patients, with policies and procedures in writing. Further, hospitals must identify all patients who are likely to suffer adverse health consequences after discharge in the absence of adequate discharge planning. Hospitals also are required to provide for the initial implementation of the discharge plan. Despite Medicare’s evident interest in promoting high-quality transitional care, the program currently lacks vigorous monitoring and reimbursement incentives.

Beyond an appropriate alignment of “sticks and carrots,” improvements in transitional care quality will require research advances, inculcation, and cultural change within health care organizations and professions, and discipline-specific training. In that spirit, this chapter is intended to provide didactic material for training health care providers who are responsible for managing the care transitions of elders with complex care needs. Much of the data presented and recommendations made apply to transfers from acute hospitals, but the key principles apply to transitions involving other health care settings as well.

With increasing pressures to shorten hospital lengths of stay, it is becoming less common for an older patient to be cared for in the hospital until she feels well, at which time she is discharged home in the care of her family. National statistics from 2004 for Medicare patients discharged from acute hospitals show that about a quarter of them were transferred to another institution (such as a nursing home or rehabilitation facility), and an additional 14% were discharged to their home with the involvement of home health professionals. Recent work also confirms many providers’ impressions that the period immediately following hospital discharge can be characterized by continued movement from place to place. A recent study of Medicare beneficiaries examined the 30-day period following hospital discharge, and found that 60% of patients made a single transfer, 18% made two transfers, 9% made three transfers, and 4% made four or more. These frequent transfers represent discontinuities in patients’ care where vital information may be lost and care plans become fragmented.

The risk for hospital readmission is greatest in the period immediately after discharge. A study of patients with hip fracture discharged from New York City hospitals, for example, found that 32% were readmitted within a 6-month period, and over a third of these readmissions occurred within the first month after discharge. It is increasingly recognized that the “peridischarge” period is a dangerous time for patients. In one prospective cohort study of patients discharged home from the medical service of an academic medical center, 6% of patients had suffered preventable adverse events by about 3 weeks after discharge. An additional 6% experienced adverse events whose severity might have been reduced by improved quality of care. In another single-institution study of patients discharged from an academic medical service, researchers examined inpatient and outpatient records for evidence of three types of errors arising from discontinuity of care: (a) medication continuity errors, (b) test follow-up errors (in which test results were pending at the time of hospital discharge, but were not acknowledged in the outpatient chart), and (c) work-up errors (in which an outpatient test was recommended or scheduled by the inpatient team, but did not take place). About half of the patients’ charts were found to contain at least one of these discontinuity errors, and those patients with work-up errors were over six times more likely to be rehospitalized than patients without these errors.

A major cause of the peridischarge hazard may be communication failures between hospital physicians and physicians who see patients soon after discharge. Historically, the same physician attended to the patient in hospital and in clinic, but the increased use of hospitalists and covering physicians creates opportunities for information discontinuities. In one study, only 3% of primary care doctors reported discussions with hospital physicians about discharge plans. In other studies, about half of discharge summaries are received by primary care physicians by 4 weeks of discharge; a quarter of summaries never reach the primary physician. And those summaries that are received commonly lack important information. Diagnostic test results are missing in 38%, consultant recommendations in 52%, discharge medications in 21%, and test results pending at discharge in 65%. There is a pressing need to improve the fidelity of information transfer from hospitals to primary care providers and other physicians and care providers.

Poor communication between hospital providers and patients (or families) probably contributes as well to peridischarge risk. Patients often report that they do not understand potential adverse effects of their discharge medications, when they may resume normal activities, which questions to ask and whom to address them to, and which warning signs to watch for. More globally, many patients express a lack confidence in their ability to make sure the care plan reflects their particular preferences, goals, and needs.

Metrics are needed to implement a quality improvement effort in transitional care. Coleman et al. developed the Care Transitions Measure (CTM), a patient-focused instrument that can be used to track performance changes over time. Two versions of the instrument, one with three items and another with 15, can be found at www.caretransitions.org. The CTM encompasses four domains of quality in transitional care (Table 16-3). The first domain is the fidelity of information transfer. Is accurate information transferred from sending clinicians to receiving clinicians in a reliable and timely manner? The second domain is patient and family preparation. Do they understand the purpose of the new care setting? If the patient is going home, do they understand the things they are responsible for managing in the care plan? The third domain is support for patient self-management. Does the patient know the warning signs and symptoms to watch for after discharge, and the purpose and adverse effects of the medications? Finally, the fourth domain is patient empowerment to assert individual goals and preferences. Does the patient feel the hospital staff took into account her particular goals when determining what her health care needs are, and where they could best be met after discharge? Lower CTM scores have been shown to predict reduced likelihood of ED use and rehospitalization in the postdischarge period. While developed for quality improvement efforts, all clinicians would benefit from considering the CTM items or Table 16-3 when contemplating individual patient transfers. The National Quality Forum has endorsed the CTM for use in public reporting.

Experienced clinicians develop a sense for which patients are likely to experience difficulties after a transition to a new setting. In hospital jargon these patients are potential “bounce-backs” or candidates for early readmission. One statistical model that helps to identify such at-risk patients is provided in a study by Coleman et al. (Posthospital care transitions: patterns, complications, and risk identification. Health Serv Res, 2004;39(5):1449). A consensus of factors that help to identify patients at risk of hospital readmission is listed in Table 16-4. Several of the items are not surprising—older patients, those with more serious illnesses, those with need for activity-of-daily-living assistance but with marginal support, and those with a history of previous hospitalizations are all more likely to require rehospitalization soon after discharge. In addition to physical ailments, however, depression predicts readmission, as does a worse self-rating of patient health. Patients who have difficulty with treatment adherence (to medication regimens or follow-up appointments) are at risk as well. Finally, patients (or families) who have not received education around the time of discharge are more likely to suffer complicated transitions.

Knowledge of risk factors for problematic transitions is important for clinical leaders who are contemplating new programs and innovations in transitional care. Such programs may be time- and labor-intensive; administrative leaders or funding agencies will look for evidence of cost-effectiveness. A key strategy in creating a cost-effective program is to employ targeting criteria that help in identifying patients most in need of special help. Patients who lack such criteria (such as those shown in Table 16-4) are less likely to require costly interventions in order to avoid transitions-related mishaps.

A number of interventions to improve care transitions have been studied; the majority involve discharge from the acute hospital. One systematic review of 54 randomized, controlled trials involving interventions designed to improve outcomes of discharge of elders from hospitals found that the location of the intervention appeared to be an important factor. The most successful interventions, in terms of reducing readmission risk, were those that took place in the patients’ homes, or in both hospital and at home, but not exclusively in the acute setting. One such landmark trial, reported by Rich and coworkers, focused on the discharge of high-risk elders with congestive heart failure (CHF). This diagnosis is the leading cause of hospital admission for Medicare patients and is associated with frequent readmission. Rich’s multifactorial intervention included intensive education about CHF provided by nurses; social work involvement for help with discharge planning and arrangement of care after discharge; review of the drug regimen by a geriatric cardiologist; dietary assessment and education by a dietician; high-intensity follow-up by the hospital’s home care services; and additional home visits and telephone consultation as needed by the research team. Those patients in the intervention arm of this prospective, randomized, controlled trial demonstrated a lower risk of readmission, reduced cost of care, and greater quality-of-life scores.

A later study by Naylor and colleagues considered a broader range of diagnoses for hospitalized elders being discharged to the community. The study population in this prospective, randomized trial included elders with one of several medical or surgical diagnoses, and known to be at risk (using predetermined criteria) for poor discharge outcomes. The intervention made use of advanced practice nurses who assumed responsibility for transitional care from the time of admission to the hospital. These advance practice nurses followed their patients into the home following discharge, and substituted for visiting nurses during the first 4 weeks after patients left the hospital. As with the Rich study, the Naylor study found that the intervention reduced risk of readmission, as well as total cost of care.

Coleman and colleagues reported on the results of the Care Transitions Intervention, which introduced two conceptual innovations. First, this intervention was designed to encourage older patients (and their caregivers) to assert their own preferences, and to acquire self-management skills relating to transitions across care settings. Second, the intervention introduced the role of a “transition coach,” who, although a nurse practitioner by training, did not function as a health care provider. Rather, she encouraged self-management and direct communication with primary care providers. The transition coach visited patients in the hospital and at home, and also made frequent follow-up telephone calls. Patients worked with the transition coach for about 28 days after their discharge to home. Major areas of emphasis included medication self-management; use of a patient-centered record (maintained by the patient, and listing medical problems, medications, and other vital information); empowerment of the patient to schedule and complete follow-up visits with physicians; and patient’s understanding of events that suggest that a health condition is worsening. Patients who received this program were significantly less likely to be readmitted to the hospital and the benefits were sustained for 5 months after the end of the 1-month intervention. More than simply managing posthospital care in a reactive manner, the Care Transitions Intervention imparted self-management skills that proved beneficial after the end of the program. Anticipated cost savings for 350 chronically ill adults extending from an initial hospitalization over 12 months is almost $296 000. Patients who received this program were also more likely to achieve self-identified personal goals around symptom management and functional recovery.

A number of hospitalist programs have set up postdischarge follow-up clinics in which patients are seen soon after discharge by the physician who provided care in the acute setting. The benefits of these programs have not yet been demonstrated in large, prospective trials. Nevertheless, a large cohort study using an administrative data base for hospitalized patients in Ontario, Canada, found that patients who followed up after discharge with their inpatient provider had a reduced combined rate of readmission and 30-day mortality.

Several pharmacist-led transitions interventions have been recently reported. One intervention involved simply a follow-up phone call placed by pharmacists to patients 2 days after discharge, during which patients were asked whether they had obtained their medicines and understood how to take them. Patients receiving the phone call returned to the ED less often than those who did not (10% vs. 24% at 30 days). Another randomized study examined the effect of pharmacists counseling patients at hospital discharge, and calling 3 to 5 days later. Patients receiving the intervention showed a lower rate of preventable adverse drug events 1 month after hospital discharge. The process of medication reconciliation (discussed below), conducted by pharmacists, was tested in a preintervention/postintervention study involving patients who resided at a nursing home, were hospitalized, and subsequently discharged back to the nursing facility. After the intervention began a pharmacist reconciled the nursing home and hospital medication lists, and reported discrepancies to patients’ physicians. Compared to the preintervention group, the postintervention group showed significantly reduced odds of suffering discrepancy-related adverse drug events (odds ratio 0.11, p = 0.05).

Use of electronic methods to transfer patient information between various sites of care (e.g., hospitals, nursing facilities, EDs) is being piloted in a number of locations. Preliminary reports suggest these methods may improve outcomes, such as reducing the number of hospital readmissions. Although these approaches present challenges (nursing homes would need to master electronic interfaces and equipment and to use federally endorsed standards), they appear promising.

A crucial first step in making systematic improvements in transitional care is for the various parties in this activity—the team of clinicians sending a patient to a new setting, the receiving team, and the patient (and family)—to clearly understand their respective roles and responsibilities. As shown in Table 16-5, certain responsibilities belong to one party alone, but some are common or shared.